Cloud Publications International Journal of Advanced Remote Sensing and GIS 2013, Volume 2, Issue 1, pp. 160-172, Article ID Tech-102 ISSN 2320-0243 Research Article Open Access Influence of Micro-Climate Parameters on Natural Vegetation A Study on Orkhon and Selenge Basins, Mongolia, Using Landsat-TM and NOAA-AVHRR Data Murali Krishna Gurram 1, Oyuntuya Sharavjamts 2 and Nooka Ratnam Kinthada 3 1 COWI India Pvt. Ltd., Plot No. 122, Phase-I, Udyog Vihar, Gurgaon, Haryana, India 2 School of Ecology and Technology Development, Mongolian State University of Agriculture, Mongolia 3 Dept. of Geoinformatics, AdiKavi Nannaya University, Jayakrishnapuram, Rajahmundry, East Godavari, India Correspondence should be addressed to Murali Krishna Gurram, murali.krishna.gurram@gmail.com Publication Date: 10 July 2013 Article Link: http://technical.cloud-journals.com/index.php/ijarsg/article/view/tech-102 Copyright 2013 Murali Krishna Gurram, Oyuntuya Sharavjamts and Nooka Ratnam Kinthada. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Abstract A remote sensing study was carried out to map the spatial distribution and intensities of various micro-climate parameters such as surface temperature, surface albedo, precipitation etc., to assess their influence over vegetation health of the region. Remote sensing data pertaining to Landsat-TM and NOAA-AVHRR was used to retrieve the vegetation parameters like NDVI, pasture yield as well as climate parameters like surface temperature and albedo. The study has demonstrated the utility of remote sensing in monitoring the seasonal variations in vegetation growth with respect to meteorological conditions and changes occurring between different periods of same seasons due to change in micro-climate. A direct relationship between micro level variations in the vegetation growth with respect to meteorological conditions in the region was established. Keywords Micro-climate Change, NDVI, Landsat TM and NOAA/AVHRR Data 1. Introduction Variations in the vigor and pattern of the vegetation manifested over the rangelands and croplands are purely of local character (Lee et al., 2002). Proper identification, classification and mapping of these patterns of high intensive and complex nature requires frequent monitoring as the change in the vigor of the vegetation occurs within a short span of time and mostly depends upon various microclimatic and terrain conditions. Moreover, change in the vigor differs from one plant species to another. To be able to notice or track, interlink and interpret these changes, observations should be made as frequently as possible i.e., once in every 2-5 days for crops, 2 days for pasture plants on the fenced area and 10 days for pasture plants that grow in the open area. And it is impractical to use any ground based survey techniques for the mapping of the same as they are very time consuming and labor intensive. Studies conducted using remote sensing technology has been proved the best for analyzing such scenarios (Bat-Oyun and Singh, 2005). Importantly, data acquired by the remote
sensing sensors, like, NOAA-AVHRR of high temporal resolution and LANDSAT-TM of relatively high spatial resolution proved to be very effective in determining the patterns of vegetation spreads over a large territory and directly retrieving the parameters such as NDVI, surface temperature and albedo (Oyuntuya et al., 2005). Apart from that, the study is one of the supporting materials for application of remote sensing and comparison of the same results in ecological studies. 1.1. Data Used Meteorological data parameters, such as temperature and rainfall were obtained from five different stations located in and around the Orkhon-Selenge basin. Meteorological data recorded during 1961-1990 were used for the study. The following table (Table 1) lists the names and location coordinates of the various meteorological stations from where the data pertaining to meteorological parameters were collected. Table 1: List of Meteorological Stations in the Study Area S. No. Name of the Meteorological Stations Longitude Latitude 1 Darkhan 105 0 54 E 49 0 28 N 2 Sukhbaatar 106 0 12 E 50 0 14 N 3 Bulgan 103 0 33 E 48 0 48 N 4 Khutag 102 0 42 E 49 0 23 N 5 Baruunkharaa 106 0 40 E 48 0 19 N NOAA-AVHRR data of 1 km of spatial resolution pertaining to 4 different months starting from May to August between the years 1989-2011 were used for the study. Similarly, multi-temporal data by LANDSAT s Thematic Mapper (TM) sensor of 30 m spatial resolution pertaining to three different years of same season i.e., August 1996, September 2007 and July 2009 were also used for the study. Details of the satellite datasets used for the study are mentioned below. NOAA AVHRR data : 1989-2011 (May to September, average of 10 days) LANDSAT-TM data : 20 th, August 1996; 20 th, September 2007 and 24 th, July 2009 Pasture yield : Every 10 days (1989-2011) Surface temperature : Average of 10 days (1989-2011) 1.2. Study Area According to agricultural scientists Orkhon-Selenge basin in Mongolia is considered as the land for crop production as it includes all lowlands or hollows between the mountains. The study area encompasses the whole territories of Selenge, Darkhan-Uul, Orkhon and Bulgan aimags (districts) and parts of the Soums Tuv aimag (Figure 1). International Journal of Advanced Remote Sensing and GIS 161
Figure 1: Map Showing the Orkhon-Selenge basin in Mongolia 1.3. Objectives of the Study The study is aimed at mapping and assessment of various vital land surface parameters, their relationships and patterns using remote sensing in the Orkhon-Selenge basin region. To accomplish the objectives, the following scenarios were tested: Determining the relationship between land surface conditions like NDVI, surface temperature, albedo and meteorological parameters. Classification and mapping of the vegetation patterns over a large territory. Analyze the patterns of both rangeland vegetation of local character such as crops (observations should be made every 2-5 days), pasture plants on the fenced area (every 2 days) and pasture plants (every 10 days) on open area. International Journal of Advanced Remote Sensing and GIS 162
2. Methodology The study requires key data inputs pertaining to meteorology and vegetation. The methodology mentioned below provides a brief description of the steps involved in collection, processing, retrieval and analysis of the data collected from different sources. 2.1. Data Processing, Retrieval and Analysis To be able, effectively monitor and map the vegetation types of the entire region, satellite images pertaining to two different sensors of different spatial and temporal resolutions and periods have used for the study. The data acquired by the NOAA-AVHRR of very good temporal resolution, and LANDSAT-TM of relatively higher spatial resolution were used for the study. 2.1.1. NOAA AVHRR Data Processing In order to retrieve three different vital parameters such as NDVI, Surface temperature and albedo, 10 day interval temporal datasets of NOAA-AVHRR data of August-September during 1989-2011 were used. Particularly, NOAA-AVHRR data is very helpful for mapping the vegetation growth scenarios so that it can keep track of all the changes on a regular basis (Gupta et. al., 1997 & 2001). The NOAA-AVHRR data sets were geometrically corrected and standard algorithms for atmospheric corrections were applied. NDVI has been calculated as mentioned below: NDVI = (Red NIR) / (Red + NIR) Subsequently, land surface temperature and pasture yields were calculated to correlate the data with NDVI values. 2.1.2. NDVI vs Pasture Yield The NDVI values and pasture yield were retrieved to see the relationship between the two and to evaluate the correlation between them (Erdenetseseg and Erdenetuya, 2005). The NDVI values and pasture yield patterns are shown in the Figures 2 and 3 below: Figure 2: Maps Showing Seasonal Variations in NDVI from May to August International Journal of Advanced Remote Sensing and GIS 163
Figure 3: Pasture Yield Retrieved from NOAA AVHRR Data Regression analysis of NDVI and pasture yield has shown the correlation as shown in the Figure 4 below: Figure 4: Regression Curves Showing Relationship between Pasture yield & NDVI 2.1.3. NDVI vs LST An attempt was also made to see how NDVI values are correlated with land surface temperature (Azzaya et al., 1998). The following figures (Figure 5 and 6) with results corroborate the relationship between these two parameters. International Journal of Advanced Remote Sensing and GIS 164
Figure 5: Relationship between LST & NDVI (Vegetation period: from May to September) Table 1: Correlation between LST & NDVI (Vegetation period: from May to September) Month Temperature (in o C) NDVI value Correlation coefficient May to Sep. 5.1 to 33.1 0.032 to 0.752 0.338 May 7.6 to 25.8 0.032 to 0.536 0.433 Jun. 13.1 to 31.4 0.072 to 0.640 0.101 Jul. 17.8 to 33.1 0.176 to 0.752-0.184 Aug. 13.3 to 27.5 0.2000 to 0.688-0.191 Sep. 5.1 to 19.3 0.104 to 0.584 0.137 Figure 6: Regression Curves Showing the Relationship between NDVI and LST The study investigated the possible use of NOAA AVHRR data for diurnal temperature measurements and mapping of its dynamics. It is determined that the radiant temperature is high on top of high mountain regions with perennial snow cover, sandy, plain, lakes and forest. International Journal of Advanced Remote Sensing and GIS 165
2.2. Landsat TM Data Processing Landsat TM multi temporal datasets were processed to retrieve the NDVI, LST and albedo on high spatial resolution and to generate the land cover theme which is used to correlate with the meteorological and other parameters retrieved from NOAA AVHRR data. Landsat s TM data was also used keeping in mind that it helps in mapping the vegetation patterns at micro-level which is very difficult to identify using the NOAA AVHRR data which is of coarse resolution. Parameters like NDVI (Figure 7) and surface albedo (Figure 10) were retrieved using the combination of Landsat TM s VIS and NIR bands. Thermal bands (MIR) are used to calculate the land surface temperature or LST (Figure 9) according to the methodology suggested by Prasanjit Dash et al. (2001). In addition to this, various land cover classes of the region were delineated using unsupervised classification method with maximum likelihood algorithm. This also helps in the proper identification and classification of different vegetation types which can be used to correlate with other parameters used in the study. Figure 7: NDVI Derived from the Landsat-TM Table 4: Variations in the Intensity of NDVI with Respect to Land Use Land Classes Date 24-07-2009 20-09-2000 Pasture 0.12 to 0.37 0.08 to 0.26 Forest 0.41 to 0.71 0.40 to 0.70 Cultivated area -0.05 to 0.3-0.06 to 0.01 Meadow 0.47 to 0.58 0.31 to 0.45 Agriculture area 0.27 to 0.56 0.01 to 0.28 International Journal of Advanced Remote Sensing and GIS 166
24-07-2009 20-09-2000 Figure 8: Total Number of Pixels by Each NDVI Class Figure 9: Surface Temperature Classes Retrieved from Landsat-TM International Journal of Advanced Remote Sensing and GIS 167
Table 5: Surface Temperature Values ( 0 C) Shown by Different Vegetation Classes Land Cover Date 24-07-2009 20-09-2007 Pasture 36.3-49.6 27.0-30.7 Forest 27.7-38.5 25.7-27.4 Cultivated area 45.6-50.5 27.2-35.3 Meadow 27.8-36.8 23.5-27.2 Agriculture area 33.4-44.1 24.6-30.1 Lake 27.0-31.4 14.2-22.0 River 27.2-29.1 15.6-22.2 Figure 10: Mapping of Surface Albedo 3. Results and Discussion Table 6: Surface Albedo Values Shown by Different Vegetation Classes Land Cover Class Date 24-07-2009 20-09-2007 Pasture 0.098-0.128 0.095-0.135 Forest 0.115-0.166 0.115-0.165 Cultivated area 0.098-0.136 0.119-0.156 Meadow 0.165-0.203 0.128-0.203 Agriculture area 0.128-0.203 0.119-0.156 Lake 0.047-0.097 0.068-0.075 River 0.084 0.031-0.075 Analysis of meteorological data pertaining to the years 1981-2010 has indicated that the average annual air temperatures during the period was less than zero or between -0.12 0 C and -1.25 0 C, while it was -0.36 0 C at Bulgan station and warmer or 0.01 to 1.0 0 C at other stations during 1991-2006. These variations in the average annual air temperatures between 0.73 and 1.22 0 C can be attributed to the changes occurred to local terrain and feature conditions. It is noticed that, the annual average precipitation in Orkhon-Selenge basin is 339.8 mm; however it is decreased to 6.4 to 43.5 mm during 1991-2006. Variations in monthly average precipitation increased International Journal of Advanced Remote Sensing and GIS 168
slightly in all areas during the months of January, May, November and December and decreased or changed drastically during the remaining months. It is decreased by 9.3 to 15.9 mm in other areas during the warmest month of August, while it is increased by 4 to 12.4 mm in Khutag and Bulgan regions. According to the meteorological conditions observed for one year, the average NDVI values ranged between 0.381 and 0.432 (Figure 4) during vegetation growth period, and for multiple years, there is a slight reduction in NDVI. Figure 11: Meteorological Conditions vs NDVI Depending on the climatic conditions the average NDVI values varied between 0.381 and 0.432 during the vegetation growth period. However, on overall basis the annual average observations have shown a decreasing trend in the NDVI values. Air temperatures in Orkhan-Selenge basin increased to 0.73-1.22 0 C in the last 15 years. The annual average air temperatures are above zero and there is a tendency of further warming in the areas other than Bulgan. Total annual precipitation in the region is decreased by 6.4-43.5 mm and precipitation received during the warm seasons dropped to 0.6-1.3%, with total yearly precipitation accounting for 91.9-95.4%. The vegetation growth in open pastures has shown a positive trend with good or higher correlation coefficient. Based on the NDVI values the vegetation growth in the regions are assessed and mapped accordingly. It is noticed that, increased temperature, prolonged warm seasons and slight decrease in precipitation seems to be favorable for the growth of vegetation. However, such situation is not observed in the study area. Frequent occurrences of unstable climatic conditions, such as, overheating of air and soil during fragile period of plant growth, increase in the precipitation rate and decrease in the frequency of precipitation found to be the reasons for poor vegetation growth. A positive correlation was found between NDVI and surface temperatures when the temperature is less than 20 0 C and inverse correlation appeared with the increase in the temperature beyond 20 0 C. Low humidity and arid conditions with temperatures 20 0 C can be considered as unfavorable for vegetation growth. International Journal of Advanced Remote Sensing and GIS 169
Depending on the vegetation growth period NDVI values recorded as 0.08-0.37 for pasture, 0.38-0.71 for forest, 0.31-0.59 for river meadow, 0.01-0.7 for crop fields and -0.06 to 0.06 for fallow land. The temperature in a year varies with land types and is highest for fallow land (45.6-50.5 0 C) and the least for river water (27.2-29.1 0 C). On the other hand, surface albedo is found to be highest for meadow surface (0.165-0.203) and lowest for water surface (0.047-0.097). NDVI found to be high for fallow land compared to pasture, i.e., by 0.2-0.48 which can be consolidated to compare with the primary variety of succession with the biomass, which shows relatively higher productivity. There is a positive as well as medium correlation noticed between NDVI and surface albedo, while there is inverse and slight to medium correlation between surface temperature and albedo. It is appropriate to see them together for monitoring terrestrial vegetative patterns. Table 8: Table Showing the Relationships between Different Surface Characters Analysis of surface patterns during the years 2006 to 2007 and 2008 to 2009 clearly indicates that there is a degradation of vegetative cover and crop production in the river basin area. The change pattern is shown in the Figures 12A and 12B. International Journal of Advanced Remote Sensing and GIS 170
Figure 12A: Changes Identified along the River Course Areas Figure 12B: Changes Identified in the Agricultural Land Use Pattern 4. Conclusion Remote sensing based analysis highlighted the relationships and influence of land cover parameters vis-à-vis meteorological conditions on micro-climate of the region. NDVI in the open pasture areas have shown a very positive correlation trend with the vegetation suggesting it can be directly used as a measure for mapping and assessment of vegetation growth. NDVI has shown a positive correlation with surface temperatures are less than 20 0 C, and at the same time NDVI has shown an inverse correlation when the temperature is more than the 20 0 C. Under the low humidity and arid conditions, temperatures are found to be lower than 20 0 C which is considered to be unfavorable for vegetation growth. There is a positive and medium correlation found between NDVI and surface albedo. Similarly, there is an inverse and slight to medium correlation found between surface temperature and albedo. Therefore, it is very much essential to take all these parameters into consideration for monitoring the regional surface land cover or vegetation patterns. International Journal of Advanced Remote Sensing and GIS 171
References Azzaya D., et al. Land Surface Temperature (LST) Estimation Using Satellite Data NOAA-AVHRR over Mongolia. MNU. 1998. 5; 138. Bat-Oyun et al. Drought Assessment over Mongolia Using Remote Sensing and Meteorological Data. Papers in Meteorology and Hydrology. 2005. 27 (5) 59-68. Erdenetseseg D., et al. Application of NDVI/AVHRR/NOAA for the Estimation of Pasture Above- Ground Biomass. Papers in Meteorology and Hydrology. 2005. 27 (5) 34-37. Gupta R.K., et al. The Estimation of Surface Temperature over an Agricultural Area in the State of Haryana and Punjab, India, and Its Relationship with the NDVI, Using NOAA-AVHRR Data, Int. J. Remote Sensing. 1997. 18; 3729-3741. Gupta R.K, et al. Estimation and Validation of Roughness Length, Surface Temperature and Sensible Heat Flux Computed From Remote Sensing (Wifs and NOAA/AVHRR) Data, Journal Of Agrometeorology. 2001. 3 (1 & 2) 189-215. Lee R. et al. Evaluating Vegetation Phenological Patterns in Inner Mongolia Using NDVI Time-Series Analysis. Int. J. Remote Sensing. 2002. 23 (12) 2505 2512. Oyuntuya Sh., et al. The Satellite Based Estimation of Land Surface Energy. Papers in Meteorology and Hydrology, Mongolia. 2005. 27 (5) 12-20. Oyuntuya Sh., et al. Results of Land Surface Processes Using Remote Sensing Technology in Selenge Aimag, Mongolia. International Conference Natural Resources and Sustainable Development in Surrounding Regions of the Mongolian Plateau. Ulaanbaatar. 2005. 200-205. Prasanjit Dash, et al. Retrieval of Land Surface Temperature and Emissivity from Satellite Data: Physics, Theoretical Limitations and Current Methods. Journal of the Indian Society of Remote Sensing. 2001. 29 (1 & 2) 23-30. International Journal of Advanced Remote Sensing and GIS 172